Do you ever wonder who's responsible for the waveforms and sound patches in your favourite synth? If it's a Korg keyboard, then the answer might be Steve McNally, one of their team of sound developers.
Steve McNally is part of the small in‑house team at Korg who have been responsible for filling the company's keyboards with factory sounds since the M1 days, and his most recent work can be heard on the i30 keyboard. The team of around five or six sound designers includes people from around the world — an international mix makes perfect sense if an instrument is to have worldwide appeal.
Steve first worked for Korg in Canada, but is now employed as a consultant to the company in Japan, designing both sounds and accompaniment styles for a range of keyboards; and when he's not designing sounds, he's travelling the world showing them off at trade shows and demonstrations. I was curious to know at what level sound designers get involved in a product — are they confronted by a finished machine with empty patches, or do they also have a say in which PCM waveforms are built into the machine in the first place?
Steve McNally: "If we were to look at the way it worked for the Trinity, for example, it's a kind of three‑stage process. We gather PCM [sounds] from everywhere, and we know essentially what we're looking for — in addition to the meat and potatoes stuff, that is. We look for interesting things, like pot tops or whatever, then we get together and look at the material we've got. In the case of the Trinity, we had two string files where one was more classical and the other was more for pop applications, and we couldn't decide between the two of them. So we decided that we'd reduce their size somehow and use both of them. At that stage, the sounds are burned into ROM for the prototypes. Then we go back to our respective countries and make programs for guitars, pads, electric pianos or whatever, using these waveforms.
"The next stage is that we meet up again in Japan, and one day we'll do the guitar patches, another day we'll do another sound type. For example, we need a guitar, a fretless bass, and a slap bass — somebody might have something slightly unusual that's interesting, then we fight and go through the politics until we arrive at the set of sounds that will actually go into the instrument. Once we've decided on the programs, you then go back and make combinations of the patches — piano and strings, and so on — then we meet again to decide which ones to use."
Where do the more abstract sounds, such as pads that evolve or tinkle or do other strange things, come from? Do you have any examples of how these things are put together?
SM: "I've never been a programmer in the sense of saying, 'I'll take this sine wave and make a piano out of it'. I've always started out with the PCM waveforms and then tried to use them in interesting ways. Years ago, on the T‑series, I made a patch called 'Dusty Sax'. It was a combination program, but there was this fluttery pitch‑enveloped noise which I layered with the sax. Next, I pitched the fluttery sound up by a fourth, but it was a bit in the way, so I thought I'd try routing it into the reverb and then using it completely wet. That way I ended up with a sax that had this tinkly reverb about a fourth up, and it was like 'wow!'. But you never really set out with that kind of end result in mind, they just happen, and weird things are kind of my speciality. If the other guys want something weird, it's a case of, 'let's bring in Steve's stuff!'."
...just saying that something has moving filters isn't going to make people want to go out and buy it.
What kind of PCM sounds do you have to choose from, apart from the obvious instrument sounds and voices? Do you get bamboo chimes and people scraping piano strings, or what?
SM: "Sure, and we did some interesting PPG‑type things to some of them. I don't know if you remember that thing called 'Flutter' — I think it was on the X3, and it was a loop of something one of the guys had come up with from joining things together. Again in the T‑series, there's a particular sample of a pot lid, and it was just gorgeous. You could play that as a tuned instrument and it was stunning — almost an electric piano type of sound."
Pretty much every synth you buy now uses PCM samples as the raw sound elements, but in the past Korg have produced ground‑breaking instruments that worked on different principles, for example the Wavestation, and the application of physical modelling in the Prophecy and Z1. Is physical modelling likely to re‑emerge in a low end machine?
SM: "If you look at what the Prophecy cost for a monophonic synth, you can now have a 12‑voice expandable Z1, which is far more powerful, for not that much more money. The difficulty is knowing where to place these instruments. Take the Wavestation: I was in love with that thing, and I'd sit down with people and show them how to put together really interesting wave‑sequenced sounds, add vector control and velocity switching, and they'd go 'let me hear the piano'. I think that there was a period where in pop music when there were lots of great keyboard sounds around that people just weren't using. The great thing about dance is that people are more experimental with their sounds, and that's now filtering over into pop music."
But surely, even when interesting keyboard sounds weren't being used on pop records, you could still hear them everywhere on film and TV commercials. Wouldn't it be practical to take some of the more experimental elements from these older instruments and combine them with the more conventional sounds that people expect to find? Without something that's new, everything converges on the GM sound set.
SM: "When we were developing our physical modelling technology, we had this thing that could deal with virtually any type of synthesis, and the first thing we did was ask if it could do wave‑sequencing, but wave‑sequencing is pretty intensive in the amount of processing power it needs. However, I agree totally — I loved that ability to shape sounds. What happened with the Wavestation is rather interesting in that it was really just after the M1, and we wanted to make a synth that wasn't PCM playback — a kind of anti‑M1. It wasn't until later that Korg expanded the Wavestation to have pianos, guitars, drums, etc. I think that made the instrument more interesting."
Do you think that samplers have taken over the role of producing abstract sounds, or do you think there's still a need for that capability in synthesizers?
SM: "Ever since the M1 there's been the problem of deciding what to make next, and even though we might not have any ideas at the time, we seem to come up with something new to put into the next model. Guys that have samplers don't seem to be that interested in this type of instrument anyway. If they're going to do a piano track, they'll load up a 64Mb piano sample and still have room for huge, unlooped drum sounds."
I've often felt that, since the M1, PCM‑based instruments haven't advanced much, other than to sound cleaner and perhaps offer more sounds.
SM: "In some ways that's true, but the Trinity is a different story. With the flexibility of the effects, sound quality and the Physical Modelling option you get a real improvement from the typical PCM keyboard."
Do you see any advantage in adopting more complex filter types, rather than the usual resonant low‑pass synth filters, to help articulate sounds?
SM: "Yes, although putting on my marketing hat, just saying that something has moving filters isn't going to make people want to go out and buy it. The keyboard has to have more than that."
But surely these days you expect a good GM set in any instrument, almost regardless of what it is, so why not give them that plus the 'added interest' features in the same machine?
SM: "I think that's a good idea. Perhaps the Wavestation would have been even more successful if we'd had drums, piano, guitar and bass in it right from the beginning."
So what sells a keyboard now? Is it just a good set of GM sounds, or is there more to it? After all, there must be countless instruments that do a good job of delivering the standard sounds.
SM: "It's true, though I've always felt that in the past Korg have produced better sounds than the competition. These days you can walk up to any keyboard and it will sound great — and even if you don't like the piano you may think the guitar is better, so it's not so easy to say which [keyboard] is best. Still, comparing the Trinity with anything else in its price range, I think it sounds better than the other stuff."
From Tubas To Techno
When you're selling the same basic instrument around the world, how do you reconcile the different musical cultures and requirements of different countries? The Germans may want tubas and accordions where the Americans want country music sounds and we in Europe want techno.
SM: "It is very difficult, but that's one of the reasons the voicing team is international. If the voicing team was American, somebody might ask why there's only one accordion, and the answer would be, 'because we had to put it in there'. But a German sound designer might want to use five or six of them. We can put enough stuff inside a keyboard now to cover all these areas — it's really a matter of how you get to it."
Is the finite amount of sample ROM available a big problem? And if so, why don't instruments have flash ROM so that a different set of sounds can be loaded into the same machine depending on what country it's being sold in. In fact, why not produce a CD‑ROM of PCM sounds so that the end user can, in effect, choose their own PCM sound set? That way the Germans could have their 10 tubas and 20 accordions.
SM: "Occasionally, different countries load in their own programs for their market, like with the Z1 here in the UK. We've talked about this, and the concept is very attractive, but the reason you get so much from a modern machine is that the chips tend to do several jobs at the same time. If you change that so you can put in anything you want, you have to walk the fine line between the instrument being a synth and a sampler."
But surely this would combine the flexibility of a sampler with the instant patch access of a synthesizer?
SM: "It's a big job collecting that many good PCM sounds to put on a disk. It's not only a matter of collecting the sounds in the first place — you also have to loop and process them to make the samples as small as possible, so you can get a lot of different sounds into the ROM that still sound good. Then you need to make programs and combinations. It's a very big job."
Modelling: A Learning Experience
Returning to the subject of modelling, I've often said that I feel modelling works better on monophonic sounds than on polyphonic sounds. Where do you see modelling helping on polyphonic instruments?
SM: "The Z1 is pretty much what we're talking about here, and at the development meeting it was initially described as a polyphonic Prophecy. But when you actually sit down and play a Z1, although it can do all the great monophonic Prophecy stuff, that's really just the beginning, and the polyphonic sounds are very different. When you've played the electric piano on a Z1, then go back and play it on a basic PCM machine, it's apples and oranges. The way the timbre changes as you alter your playing dynamics is far more realistic. And the synth stuff is great. Once you can do FM polyphonically, it sounds so much cooler. It's like when synths became polyphonic in the old days, their capabilities soared. I think it's the same with the Z1. But you stand behind these guys at trade fairs and they stab away at a single note, and run through the patches as though they were checking out an M1. That's not how you check out the controllability of modelling. It's an educational thing where we have to encourage musicians to play these instruments in a different way, and that's the scary thing. When you sit down and play a PCM synth, everybody knows the rules, but Program B25 on the Prophecy might use a certain wheel to control the breath content of a sound, while on Program 17, the same wheel might add growl. There's no consistent user interface, so each patch is a learning experience in its own right. But that's part of the fun of it."
What would you like Korg to build next, if you had your own way, and how do you think the user interface can be improved now you're using touch screens?
SM: "I don't really know what we'll make next. As I said, we all sit down at these meetings, and wonder where we should go. I would really like to see something, along the lines of what you're suggesting, that covers the meat and potatoes thing but also has new and more interesting capabilities for changing sounds.
"I like the touch screens. I've used everything, but now I've got used to touch screens, I start poking away at an X3 screen waiting for something to happen! Some people ask for monitor outputs and mouse ports, and I guess there is that, but for me, touching a screen is faster than working with a mouse."